The differentiation of the oligodendrocyte from its bipotential progenitor culminates in the production of the myelin-specific proteins and the elaboration of membrane processes that ensheath the axon. Mutations in proteolipid protein (PLP) and its alternatively spliced isoform DM-20, the major protein constituents of central nervous system myelin, are characterized by a significant reduction in the number of mature oligoden-drocytes, resulting in severe hypomyelination, tremor and early death. The canine shaking pup carries such a mutation, a single base change that substitutes a proline for a histidine near the first transmembrane region of PLP and DM-20. This mutation hinders oligodendrocyte differentiation, as evidenced by a splicing pattern at the PLP locus characteristic of immature oligodendrocytes. The spliced transcript expressed earliest in development, DM-20, continues to be overexpressed in shaking pup oligodendrocytes. The disruption of the normal maturation schedule in these X-linked dysmyelinating disorders suggests that PLP or DM-20 plays a fundamental role in oligodendrocyte development. We propose that, while the more abundant PLP is the primary structural component of myelin, DM-20 may be critical to oligodendrocyte maturation.

Myelination is a developmentally controlled process whereby neuronal axons are enwrapped by a multilayer membrane sheath. In the central nervous system (CNS), this function is carried out by the oligodendrocyte, a terminally differentiated glial cell that sends out membrane processes to ensheath numerous axons. The assembly of the myelin sheath caps the development of the oligodendrocyte from a bipotential progenitor cell that will differentiate into either an oligodendrocyte or a type II astrocyte depending on the signals it receives (reviewed by Raff et al. 1989; Dubois-Dalcq and Armstrong, 1990). In addition to the several growth factors that have been shown to play a role in this developmental pathway, it has been suggested that proteolipid protein (PLP), the predominant structural protein in the myelin sheath, may be important in the maturation of the oligodendrocyte (Hudson et al. 1987).

PLP is an extemely hydrophobic transmembrane protein that is highly conserved between species. Alternative splicing within exon 3 of the PLP gene generates a less abundant isoform, DM-20, that is identical to PLP with the exception that it lacks 35 amino acids in the middle of the protein (Nave et al. 1987a). DM-20 is present before PLP in the developing CNS, but is gradually superceded by PLP, which predominates in mature and actively myelinating oligodendrocytes (Kronquist et al. 1987; Van Dorssalaer et al. 1988; Gardinier and Macklin, 1988). The evidence that PLP and/or DM-20 may play a role in the maturation of the oligodendrocyte comes primarily from the study of animals carrying mutations in the PLP gene. To date, mutations in the X-linked PLP gene have been identified in the mouse and rat, as well as in three distinct lineages of the human Pelizaeus-Merz-bacher disease. The jimpy mouse carries a point mutation in a splice junction that results in the deletion of exon 5 from the PLP/DM-20 mRNA, with a resulting frameshift and premature termination (Nave et al. 198lb-, Macklin et al. 1987a). In contrast, the mutations identified in the myelin synthesis deficient (jimpymsd) mouse, the myelin deficient (md) rat and the three Pelizaeus-Merzbacher lineages have all been point mutations resulting in different single amino acid substitutions present in both PLP and DM-20 (Gencic and Hudson, 1990; Boison and Stoffel, 1989; Gencic et al. 1989; Hudson et al. 198)a;Trofatter et al. 1989; Simons and Riordan, 1990). Despite the variation in the mutations detected, the phenotypes of the different mutants are strikingly similar. These mutations result in severe dysmyelination of the CNS, tremor and early death. There is a reduction in the number of mature oligodendrocytes and consequently the levels of all myelin proteins are greatly reduced (Matthieu et al. 1973; Meier and Bischoff, 1975; Skoff, 1976; Billings-Gagliardi et al. 1980a; Wolf et al. 1983; Kerner and Carson, 1984; Jackson and Duncan, 1988). This paucity of mature oligodendrocytes is not observed in animals carrying mutations in other myelin proteins, which suggests that it is not simply due to the lack of myelination (reviewed by Campagnoni and Macklin, 1988).

In this report, we describe a mutation in the PLP gene of the dog and provide molecular evidence for a role of PLP and/or DM20 in oligodendrocyte maturation. The shaking pup arose spontaneously in a line of springer spaniels and is characterized by a severe tremor beginning at about twelve days of age followed by late onset convulsions and, under natural conditions, an early death by about 3-4 months of age (Griffiths et al. 1981a). The peripheral nervous system is normal, but the CNS exhibits severe hypomyelination and a reduction in the number of mature oligodendrocytes (Duncan et al. 1983). The residual myelin present in the CNS is poorly compacted and displays several features that appear in myelin from younger control animals, such as increased amounts of cytoplasm in the lateral loops, lateral loops that terminate outward and intemodal or paranodal pockets of oligodendrocyte cytoplasm (Griffiths et al. 1981b). Previous work has demonstrated that all of the CNS myelin proteins assayed in the shaking pup are reduced in amounts, with PLP the most drastically affected, down to 0.8 % of the normal dog (Inuzuka et al. 1986; Yanagisawa et al. 1987). This result, in combination with the X-linked inheritance of the defect and the parallels between the shaking pup disorder and the jimpy mouse, suggested the PLP gene as the target for the mutation in the shaking pup. Indeed, our analysis of the PLP genes from normal dog and shaking pup has revealed a point mutation in the PLP gene that creates a histidine to proline change in both the PLP and DM-20 amino acid sequence. The identification of a canine dysmyelinating model for the human Pelizaeus-Merzbacher disease enables the further analysis of disease parameters not readily assayable in rodents (e.g. nerve conduction velocities) and more importantly, presents a better paradigm for the evaluation of therapeutic strategies for this debilitating disease. In addition, evidence is presented that indicates that the oligodendrocytes of the shaking pup are delayed in their maturation, as the developmentally regulated pattern of alternative splicing in this mutant is altered and reflects those observed in immature oligodendrocytes. In light of the initial preeminence of DM-20 over PLP in the early stages of normal oligodendrocyte differentiation, we postulate that the developmental delay in shaking pup oligodendrocytes originates from the loss of functional DM-20 protein.

Northern and Southern blot hybridization

Total RNA was isolated from brains and spinal cords by the guanidinium isothiocyanate method as described by Maniatis et al. (1982). RNA was run on 1.0% agarose gels, blotted onto nitrocellulose and hybridized to 32P-labelled probes as described by Nadon et al. (1988). Quantitation of RNA levels was accomplished by directly counting the radioactivity on the Northern blots using an Arnbis Radioanalytic Imaging System (Ambis Systems Inc., San Diego CA). Southern blots were prepared as described by Southern (1975) and hybridized overnight at 42° in a buffer containing 50% formamide (omitted for oligonucleotide probes), 5xSSPE, 5x Denhardts, 1% SDS and 100 μg ml-1 denatured salmon sperm DNA. Northern and Southern blots were washed post-hybridization in lxSSC-1% SDS at room temperature for 20-30min and in 0.1% SSC-1% SDS at 55-65° for 10-30min. The human PLP cDNA (Puckett et al. 1987), mouse MBP cDNA (de Ferra et al. 1985) and mouse GFAP cDNA (Lewis et al. 1984) were radiolabelled using the BRL Nick Translation kit (BRL, Gaithersburg MD) according to the manufacturers specifications. Oligonucleotides were synthesized on an Applied Biosystems DNA Synthesizer and were end-labelled with [32P]y-ATP (Amersham, Arlington Heights IL) in a reaction containing 0.2 mg oligonucleotide, 100μCi y-ATP, 10mM DTT, 50 mM glycine pH 9.5, 10 mM MgC12, and 5-10 units T4 polynucleotide kinase (New England Biolabs, Beverly MA). After a 30min incubation at 37°, the labelled oligonucleotides were purified by passage over a Sep-pak C18 cartridge (Waters Associates, Milford MA).

Library preparation and DNA sequencing

Genomic DNA from shaking pup and normal dog was partially digested with Sau3A restriction enzyme (New England Biolabs) and ligated into the EMBL 3 vector (Stratagene, La Jolla CA) according to the manufacturers protocol. The ligated DNA was packaged using the Gigapack Plus system (Stratagene). Phage clones were mapped by hybridization of Southern blots to exon-specific oligonucleotide probes and by lambda-Cos mapping as described by Rackwitz et al. (1984). EcoRI fragments containing PLP exons were subcloned into pT/13 plasmid vectors obtained from Pharmacia (Piscataway NJ).

The PLP protein-encoding and flanking sequences were sequenced from both the shaking pup and a normal male littermate. Sequencing of the shaking pup gene was performed on plasmid subclones of the phage genomic clones, using the GemSeq K/RT system (Promega, Madison WI) according to the manufacturer’s instructions. Sequencing of the normal dog locus was accomplished by amplification of the exons directly from genomic DNA using the polymerase chain reaction (PCR) (Saiki et al. 1988) and sequencing the amplified DNA. The sequence of exon 2 from the shaking pup was also verified by sequencing PCR amplified genomic DNA. PCR was performed with Taq DNA polymerase from Stratagene, using conditions recommended by the manufacturer. The amplification consisted of 30 cycles of a lmin denaturation at 95°, a 2 min annealing step at 55°, and a 3 min elongation step at 70° with a 7 min extension of the elongation step after the 30th cycle. PCR amplified DNA was purified by four washes through a Centricon 30 filter (Amicon, Danvers MA) before using it for sequencing. Sequencing of the PCR amplified DNA was performed according to Higuchi et al. (1988) using P-labelled primers and the Sequenase DNA polymerase (U. S. Biochemical, Cleveland OH), which is described by Tabor and Richardson (1987). Sequencing reactions were run on 6% acrylamide/8M urea sequencing gels in Tris-borate buffer, which were fixed for 30 min in 10 % methanol/10% acetic acid and dried before autoradiography. All DNA was sequenced from both strands.

The 17-mer primers used for sequencing the canine PLP gene corresponded to intronic sequences (except for the 5’ flanking region, first and last exons) as follows:

5’ flanking (antisense) 5’ TTTGTTCAGCTGGAAGG 3’, 5’GCCTGCTTGAATCTTCC 3’

Exon 1 (sense) 5’ TGTCAATCAGAAAGCCC 3’

Exon 1 (antisense) 5’ ACCATTGGAAACCCCAA 3’

Exon 2 (sense) 5’ CCACAGAGAGGTATGAG 3’

Exon 2 (antisense) 5’ TCTATATGTCTTCAGGG 3’

Exon 3 (sense) 5’ GAAGGGAACTGTCCTCA 3’

Exon 3 (antisense) 5’ CAGACTCACGCCCAATT 3’

Exon 4 (sense) 5’ TCAATGTCTACAGGCCA 3’

Exon 4 (antisense) 5’ AGTGCTTT CATAGGAGG 3’

Exon 5 (sense) 5’ AATCTCCATGGAGCCCA 3’

Exon 5 (antisense) 5’ AAAGGCCATGGGTAGGA 3’

Exon 6 (sense) 5’ GCTGCATGCATGATCTA 3’

Exon 6 (antisense) 5’ TTCCCAGGTGCTTCTCT 3’

Exon 7 (sense) 5’ TCCCAAAAGCTTTGGAG 3’

Exon 7 (antisense) 5’ CGTCAAGTAAGAAGAGG 3’

RNAse protection assay

Antisense RNA probes were made using the Gemini II Riboprobe kit (Promega, Madison WI), utilizing 200ng of PCR-generated template DNA that included the T7 polymerase site at the 3’ end. This template was made from genomic DNA using a 24 base sense primer that started 14 bases 5’ to exon 3 (5’ ATCTGTTAATGCAGGATCCATGCC 3’) and a 46 base antisense primer that included 26 bases from intron 3 (46-71 bases 3’ of exon 3) and 20 bases that encode the T7 polymerase binding site (5’ TGTAATACGACTCACTATA GATGAGGCCACAGACTCACGCCCAATT 3’). The total size of the probe produced from this DNA is 347 bases. RNAse protection assays were performed essentially as describe by Winter et al. (1985), with the following modifications. Hybridizations containing 106ctsmin-1 of probe and 30 μg total brain RNA were heated to 65° for 5’ and then incubated at 46° for 16-18 h. Digestion with 500 units mf1 RNAse T1 (BRL, Gaithersburg MD) was carried out at 16° for one hour. Samples were run on 6% sequencing gels and the amounts of radioactivity in the PLP and DM-20 bands were directly quantitated on an Ambis Radioanalytic System. The ratio of DM-20 to PLP RNA was calculated from the average of three different RNAase protection gels, after correcting the counts in the DM-20 band for the difference in size of the protected fragments.

The PLP gene from the shaking pup does not contain any gross deletions or rearrangements

Genomic DNA from a shaking pup and a normal male littermate were digested with restriction enzymes and analyzed by Southern blot hybridization using a human PLP cDNA probe (data not shown). The pattern of bands displayed by the shaking pup DNA was indistinguishable from that of the normal dog, indicating that there are no gross deletions or rearrangements in the PLP gene. In order to analyze the genomic structure of the canine PLP locus in more detail, a library was made from shaking pup genomic DNA. Phages containing inserts from the PLP locus were isolated by plaque hybridization using the human PLP cDNA, and overlapping clones were obtained. These clones were mapped by restriction enzyme digestion and hybridization to exon-specific probes. The map of the canine PLP locus is presented in Fig. 1. The gene contains seven exons, spanning a distance of 18 kb. The genomic organization of the canine PLP gene is very similar to that of the mouse (Macklin et al. 19876; Gencic and Hudson, 1990) and human (Diehl et al. 1986; Hudson et al. 1989a) PLP genes. The intron/exon boundaries are identical and only minor differences in intron size are observed between species.

Fig. 1.

Map of the canine PLP genomic locus. The structure of the canine PLP genomic locus was determined by mapping the exons within three overlapping genomic DNA clones from the shaking pup. The phage DNA was digested with BamHI (B), Eco RI (R) or both enzymes simultaneously, run on a 0.7% gel and Southern blotted. The blots were hybridized to 32P-labelled exon-specific oligonucleotide probes. Verification of the exon positions was obtained through the sequence analysis, which localized restriction sites within the exons or in the immediately flanking intron sequences. The gene spans approximately 18 kb. The exons are numbered above the line and filled areas denote protein-encoding sequences.

Fig. 1.

Map of the canine PLP genomic locus. The structure of the canine PLP genomic locus was determined by mapping the exons within three overlapping genomic DNA clones from the shaking pup. The phage DNA was digested with BamHI (B), Eco RI (R) or both enzymes simultaneously, run on a 0.7% gel and Southern blotted. The blots were hybridized to 32P-labelled exon-specific oligonucleotide probes. Verification of the exon positions was obtained through the sequence analysis, which localized restriction sites within the exons or in the immediately flanking intron sequences. The gene spans approximately 18 kb. The exons are numbered above the line and filled areas denote protein-encoding sequences.

The shaking pup PLP locus contains a point mutation in exon 2

The sequence of the normal canine PLP locus is presented in Fig. 2. Of the 276 amino acids in the PLP protein, there is only one amino acid difference between the canine and mouse, rat and human proteins, an isoleucine at nucleotide position 591 of the canine locus instead of the valine in the other species. At the nucleotide level, there is 97% conservation between canine protein-encoding sequences and those of the rat, mouse and human (reviewed by Hudson and Nadon, 1990).

Fig. 2.

Sequence of the canine PLP gene. The sequence of the normal canine PLP gene is presented. Sequences in the RNA encoding exons are in capital letters, with the numbering corresponding to the presumptive nucleotide position in the RNA, based on analogy with the human gene. Flanking DNA from the 5’ region and intron sequences are represented in lower case letters.

Fig. 2.

Sequence of the canine PLP gene. The sequence of the normal canine PLP gene is presented. Sequences in the RNA encoding exons are in capital letters, with the numbering corresponding to the presumptive nucleotide position in the RNA, based on analogy with the human gene. Flanking DNA from the 5’ region and intron sequences are represented in lower case letters.

Fig. 3 presents the sequence of exon 2 from the shaking pup. There is a point mutation at position 219 of the coding sequence that results in a histidine to proline change in the protein. This is the only sequence difference between the normal dog and the shaking pup. All protein-encoding sequences, splice junctions, 200 bases of the 3’ untranslated region, the entire 5’ untranslated region and 170 bases of 5’ flanking DNA were examined. This change in the genomic DNA was verified by virtue of the fact that the mutation creates a new Avail site in the shaking pup. Exon 2 was amplified by PCR from both shaking pup and control genomic DNA, and this PCR DNA was digested with Avail (Fig. 4). The new Avail site results in the cleavage of the PCR DNA from the shaking pup, but not from the control.

Fig. 3.

Sequence analysis of the shaking pup PLP gene demonstrates a point mutation in exon 2. The sequence of exon 2 of the PLP locus from the shaking pup is presented. There is a single nucleotide difference from the normal dog (A→C) at position 219 (in bold-face), which results in a histidine to proline change in the protein. The lanes from the sequencing gels showing the difference in the nucleotide sequence are presented in the lower panel. This is the only sequence difference detected between the normal dog and shaking pup PLP genes.

Fig. 3.

Sequence analysis of the shaking pup PLP gene demonstrates a point mutation in exon 2. The sequence of exon 2 of the PLP locus from the shaking pup is presented. There is a single nucleotide difference from the normal dog (A→C) at position 219 (in bold-face), which results in a histidine to proline change in the protein. The lanes from the sequencing gels showing the difference in the nucleotide sequence are presented in the lower panel. This is the only sequence difference detected between the normal dog and shaking pup PLP genes.

Fig. 4.

The point mutation in the shaking pup creates a new 4vaII restriction site. Exon 2 was amplified from genomic DNA of the shaking pup (SH P) and a normal male littermate (CON) by PCR using primers in the intron sequences immediately flanking the exon. The DNA was digested with Avail (recognition site: GGACC), and both digested and undigested DNA was run on a 3 % Nusieve agarose/1% Seakem ME agarose gel, using HaeIII digested ϕX174 DNA as a marker (M). Xvall cleaves exon 2 DNA from the shaking pup, but not from the control normal dog. The faint bands result from nonspecific priming during the PCR. The PCR reactions used for the Avail digests were prepared separately from the reactions used for sequencing.

Fig. 4.

The point mutation in the shaking pup creates a new 4vaII restriction site. Exon 2 was amplified from genomic DNA of the shaking pup (SH P) and a normal male littermate (CON) by PCR using primers in the intron sequences immediately flanking the exon. The DNA was digested with Avail (recognition site: GGACC), and both digested and undigested DNA was run on a 3 % Nusieve agarose/1% Seakem ME agarose gel, using HaeIII digested ϕX174 DNA as a marker (M). Xvall cleaves exon 2 DNA from the shaking pup, but not from the control normal dog. The faint bands result from nonspecific priming during the PCR. The PCR reactions used for the Avail digests were prepared separately from the reactions used for sequencing.

PLP/ DM-20 mRNA levels are greatly reduced in the shaking pup

Northern blot analysis was performed on total RNA from brain and spinal cord of shaking pups and normal male fittermates (Fig. 5A). The normal dog displays a single band of 3.0 kb, which contains the transcripts for both PLP and DM-20. By comparison, the mouse produces a major transcript of 3.2 kb and a minor one of 2.4 kb, through the use of alternate polyadenylation sites. In the shaking pup, the PLP mRNA is of the appropriate size, but is reduced to approximately 20% of control levels (Fig. 5A). The RNA for MBP, the other major CNS myelin protein, is similarly reduced to about 25 % of control levels in both brain and spinal cord (Fig. 5B). The reduction in the PLP and MBP message levels was also quantitated with slot blots, yielding the same results (data not shown). The mRNA for GFAP, an astrocyte-specific protein, is increased by 30% over control levels, possibly indicating astrocytic hypertrophy in the shaking pup (Fig. 5B). Such hypertropy is also observed in the jimpy mouse (Skoff, 1976).

Fig. 5.

Northern blot analysis of shaking pup and normal dog RNA. (A) Northern blots containing 10 μg/lane total RNA were hybridized with nick translated human PLP cDNA probe. RNA was from brains and spinal cords of two and ten week old shaking pups (S) and normal male littermates (C), and from mouse brain (Mu). (B) Northern blots containing 10μg/lane total RNA were hybridized to 32P labelled probes made from the mouse MBP cDNA or mouse GFAP cDNA. RNA was from brains and spinal cords of a ten week old shaking pup (S) and a normal male littermate (C).

Fig. 5.

Northern blot analysis of shaking pup and normal dog RNA. (A) Northern blots containing 10 μg/lane total RNA were hybridized with nick translated human PLP cDNA probe. RNA was from brains and spinal cords of two and ten week old shaking pups (S) and normal male littermates (C), and from mouse brain (Mu). (B) Northern blots containing 10μg/lane total RNA were hybridized to 32P labelled probes made from the mouse MBP cDNA or mouse GFAP cDNA. RNA was from brains and spinal cords of a ten week old shaking pup (S) and a normal male littermate (C).

DM-20 transcripts are preferentially expressed in the shaking pup

The Northern blots presented in Fig. 5 do not differentiate between PLP and DM-20 mRNAs, as they differ by only 105 bases. PLP mRNA is derived from exons 1-7 (in their entirety) of the PLP gene, while DM-20 mRNA arises when a donor splice site located within exon 3 is used in conjunction with the acceptor splice site of exon 4 (Nave et al. 1987a). To quantitate the levels of the two specific messages, RNAse protection assays were performed using an RNA probe that encompasses the alternatively spliced exon, exon 3. Fig. 6 shows the results of one such assay in which the exon 3 RNA probe was hybridized to total brain RNA from four week old and one day old shaking pups and age-matched controls. The upper band is the protected fragment resulting from hybridization of the probe to the PLP mRNA, and the lower band is the protected fragment from the DM-20 message. In normal dogs, the DM-20 to PLP ratio is 0.50 at one day of age, and decreases to 0.32 by four weeks. In the shaking pup, the ratio of DM-20 to PLP also falls as the animal ages, but it is still much higher than in controls. In fact, the ratio of DM-20 mRNA to PLP mRNA in the four week shaking pup (0.59) is very similar to the ratio observed in the one day control (0.50), and in the one day shaking pup the ratio of DM-20 to PLP mRNA is much higher (2.33), with DM-20 mRNA predominating over PLP.

Fig. 6.

The ratio of the DM-20 and PLP mRNAs is altered in the shaking pup. RNAse protection assay was performed on total brain RNA from four week old and one day old control (C) and shaking (S) pups. The 347 base probe was from the genomic sequence of the canine PLP locus, encompassing the entire third exon, 14 bases of intron 2 and 71 base of intron 3. The larger protected fragment of 262 bases corresponds to the PLP mRNA and the smaller one (157 bases) to the DM-20 mRNA. Radioactivity was quantitated on an Ambis Radioanalytic System. The marker (M) is end-labelled MspI digest of pBR322.

Fig. 6.

The ratio of the DM-20 and PLP mRNAs is altered in the shaking pup. RNAse protection assay was performed on total brain RNA from four week old and one day old control (C) and shaking (S) pups. The 347 base probe was from the genomic sequence of the canine PLP locus, encompassing the entire third exon, 14 bases of intron 2 and 71 base of intron 3. The larger protected fragment of 262 bases corresponds to the PLP mRNA and the smaller one (157 bases) to the DM-20 mRNA. Radioactivity was quantitated on an Ambis Radioanalytic System. The marker (M) is end-labelled MspI digest of pBR322.

To summarize, the RNA analysis illustrates two facets of the development of the oligodendrocytes in the shaking pup. One is that the amount of RNA from the PLP gene, and the MBP gene as well, is reduced at all ages, reflecting the paucity of oligodendrocytes (Fig. 5). The other is that although the ratio of DM-20 to PLP message decreases with age in the mutant as it does in the control, the ratio is consistently much higher in shaking pup oligodendrocytes than in age-matched controls (Fig. 6), indicating that the mutant oligodendrocytes lag behind their control counterparts in their development.

PLP is believed to play an essential role in the compaction of the myelin sheath, possibly through homophilic interactions between domains of the protein localized on the external surface of the myelin membrane (Hudson et al. 1989b). This protein exhibits a very high degree of conservation between species, with rat, mouse and human proteins sharing 100% amino acid identity (reviewed by Hudson and Nadon, 1990), and the dog diverging by only one amino acid (Fig. 3). Previous studies on other animal mutants and the human Pelizaeus-Merzbacher disease have demonstrated that even relatively conservative amino acid changes in the protein destroy its ability to function (reviewed by Hudson and Nadon, 1989). We have documented a point mutation in the dog, which produces a histidine to proline substitution, that results in such a loss of function. This was the only sequence difference found between the shaking pup and normal dog. It is extremely unlikely that the exon 2 mutation is a secondary mutation arising due to the loss of selective pressure for the proper sequence. The shaking pup was discovered quite recently (first reported by Griffiths et al. 1981n) and there have only been three generations between its discovery and the animals used in this molecular analysis. Therefore, this single amino acid substitution appears to have a great affect on the function of the protein.

It is perhaps significant that the mutation in the shaking pup results in the substitution of a proline for a histidine on the intracellular edge of the first transmembrane region of the protein. Prolines have been shown to have the capability of changing the direction of the amino acid chain, and in fact are generally excluded from intramembrane regions of proteins (Brandl and Deber, 1986). The histidine to proline substitution may result in the disruption of the alpha-helical structure of the protein chain and thus prevent the correct folding and/or insertion of the protein into the membrane. An observation that is consistent with this explanation is that oligodendrocytes in the shaking pup are characterized by prolific distended rough endoplasmic reticulum (Duncan et al. 1987). PLP is normally synthesized in the RER and transported to the membrane through the Golgi apparatus (Colman et al. 1982; Roussel et al. 1987). The mutant PLP/DM-20 protein of the shaking pup must also traverse this pathway, since PLP/DM-20 is immunocytochemically detectable in myelin sheaths (Yanagisawa et al. 1987). However, the mutant protein may assume an abnormal configuration that retards its trafficking from the RER to the Golgi and consequently impairs the transport of other proteins, resulting in the distention of the RER.

The diminished level of PLP transcripts in shaking pup (Fig. 5A) probably reflects the reduced number of mature oligodendrocytes in this mutant (Duncan et al. 1983), as a similar reduction exists for MBP transcripts (Fig. 5B). Moreover, by combining in situ hybridization histochemistry with quantitative image analysis, Verity et al. (1990) determined that the reduced PLP and MBP mRNA expression in jimpy mice was due primarily to fewer cells expressing these transcripts. It is worth noting that both PLP and MBP protein levels of shaking pup exhibit a much greater reduction (to 0.8% and 3-6% of controls, respectively (Yanagisawa et al. 1987)) than their message levels (20% and 25% of controls; Fig. 5). While it is anticipated that a mutant protein (PLP) would be readily degraded, the observation that MBP is also severely depressed suggests that oligodendrocytes have translational or post-translational mechanisms for preventing the accumulation of myelin proteins whenever myelin assembly is blocked. A discrepancy between the levels of myelin messages and myelin proteins has also been observed in other PLP mutants (reviewed in Hudson and Nadon, 1990), again indicating that oligodendrocytes prevented from fully differentiating and assembling a myelin sheath either limit the translation of myelin proteins or greatly enhance the degradation of these surplus proteins.

One of the most perplexing aspects of the PLP mutants is how a point mutation in a structural protein such as PLP could have such a dramatic effect on the fate of the oligodendrocytes. Not only is there a depletion of mature oligodendrocytes in the shaking pup (Duncan et al. 1983), but the surviving oligodendrocytes appear immature by both morphological (Griffiths et al. 19816) and molecular (Fig. 6) criteria. It is unlikely that the lack of mature oligodendrocytes in each of these disorders is due to toxicity of the mutant PLP/DM-20, since jimpy oligodendrocytes can flourish and synthesize myelin, albeit abnormal myelin, when the cells are cultured in medium conditioned by normal astrocytes or transplanted into a non-jimpy host (Gumpel et al. 1987; Bartlett et al. 1988). These experiments suggest that despite the continued production of mutant PLP/DM-20 protein, jimpy oligo-dendrocytes can survive and attempt myelination when presented with the proper factors.

Analysis of developmentally regulated protein isoforms has provided molecular evidence that the oligodendrocytes present in the shaking pup and the other PLP mutants are prevented from fully maturing. DM-20 is expressed before PLP in the human, rat, cow and mouse, and the ratio of DM-20 to PLP decreases as PLP expression rises (Kronquist et al. 1987; Van Dorssalaer et al. 1988; Gardinier and Macklin, 1988). In both jimpymsd and the shaking pup, approximately equal amounts of PLP and DM-20 protein are present during the period of active myelination, while in normal animals the ratio of DM-20 to PLP is 0.25 and 0.06 respectively (Yanagisawa et al. 1987; Gardinier and Macklin, 1988). The mutant DM-20 protein may be in excess in these two disorders because it is more stable than the mutant PLP. However, analysis of the transcripts encoding DM-20 and PLP in the shaking pup suggests that regulation occurs at the transcriptional level as well, and that the alternative splicing of the PLP locus in shaking pup oligodendrocytes is locked into the early developmental pattern observed in immature oligodendrocytes (Fig. 6). The DM-20 to PLP mRNA ratio in the four week old shaking pup is very close to the ratio observed for the one day control dog, and in the one day shaking pup, the DM-20 mRNA is even higher, accounting for 70% of the RNA from the PLP gene. Although our present work does not contain time points early enough in development to see expression of DM-20 alone, the dog does appear to follow the same pattern of oligodendrocyte differentiation observed in other species, in that the DM-20 to PLP ratio declines with age (Fig. 6; Results). Examination of the MBP isoforms that are also developmentally regulated in oligodendrocytes (Inuzuka et al. 1986) provides further support for the concept that oligodendrocyte development is retarded in the PLP mutants. The ratios of the MBP isoforms in both the jimpy mouse and the shaking pup are shifted to mirror an earlier stage of development (Campagnoni et al. 1984; Inuzuka et al. 1986).

In summary, mutations in PLP and DM-20 prevent the normal schedule of maturation of oligodendrocytes and therefore one or both of these proteins must play a role in the development of these cells. We propose that DM-20 is involved in oligodendrocyte maturation, since it precedes PLP in development. Experimental support for the suggestion that DM-20 may be an important component of the oligodendrocyte developmental pathway comes from attempts to express each isoform individually in vivo. Transgenic mice were made that carry a construct encoding the human PLP protein, but not DM-20 (Nadon et al. 1989). One line of mice that produces high levels of transgene-specific RNA was crossed to the jimpy line, and male offspring carrying both the mutation and the transgene were analyzed by immunocytochemistry and electron microscopy. There was no evidence of expression of the transgene PLP protein in these mice and they exhibited the same degree of hypomyelination as jimpy mice. Although there are several possible explanations for this result, it is consistent with a regulatory role for DM-20. If DM-20 is essential for the development of the oligodendrocytes, the cells in the jimpyμransgenic mice would still be inhibited in their development since the transgene does not make DM-20. Analysis of transgenic mice expressing only the DM-20 isoform should provide direct evidence for the function of this alternatively spliced product of the PLP locus.

The authors thank Dr N. Cowan for providing the GFAP cDNA, Pete Kelly for supplying synthetic oligonucleotides and Drs Heinz Arnheiter and Bryn Watkins for critical review of the manuscript. This work was supported in part by grants to I. D. D. from the National Institutes of Health (NS23124) and from the National Multiple Sclerosis Society (RG1791).

Bartlett
,
W. P.
,
Knapp
,
P. E.
and
Skoff
,
R. P.
(
1988
).
Glial conditioned medium enables jimpy oligodendrocytes to express properties of normal oligodendrocytes: production of myelin antigens and membranes
.
Glia
1
,
253
259
.
Billings-Gagliardi
,
S.
,
Adcock
,
L. H.
,
Schwing
,
G. B.
and
Wolf
,
M. K.
(
1980a
).
Hypomyelinated mutant mice. II. Myelination in vitro
.
Brain Res
.
200
,
135
150
.
Boison
,
D.
and
Stoffel
,
W.
(
1989
).
Myelin-deficient rat: a point mutation in exon III (A—>C, Thr75—>Pro) of the myelin proteolipid protein causes dysmyelination and oligodendrocyte death
.
EMBO J
.
8
,
3295
3302
.
Brandl
,
C. J.
and
Deber
,
C. M.
(
1986
).
Hypothesis about the function of membrane-buried proline residues in transport proteins
.
Proc. natn. Acad. Sci. U.S.A
.
83
,
917
921
.
Campagnoni
,
A. T.
,
Campagnoni
,
C. W.
,
Bourre
,
J. M.
,
Jacque
,
C.
and
Baumann
,
N.
(
1984
).
Cell-free synthesis of myelin basic proteins in normal and dysmyelinating mutant mice
.
J. Neurochem
.
42
,
733
739
.
Campagnoni
,
A. T.
and
Macklin
,
W. B.
(
1988
).
Cellular and molecular aspects of myelin protein gene expression
.
Mol. Neurobiol
.
2
,
41
89
.
Colman
,
D. R.
,
Kreibich
,
G.
,
Frey
,
A. B.
and
Sabatini
,
D. D.
(
1982
).
Synthesis and incorporation of myelin polypeptides into CNS myelin
.
J. Cell Biol
.
95
,
598
608
.
De Ferra
,
F.
,
Engh
,
H.
,
Hudson
,
L.
,
Kamholz
,
J.
,
Puckett
,
C.
,
Molineaux
,
S.
and
Lazzarini
,
R. A.
(
1985
).
Alternative splicing accounts for the four forms of myelin basic protein
.
Cell
43
,
721
727
.
Diehl
,
H.-J.
,
Schaich
,
M.
,
Budzinski
,
R.-M.
and
Stoffel
,
W.
(
1986
).
Individual exons encode the integral membrane domains of human myelin proteolipid protein
.
Proc. natn. Acad. Sci. U.S A
.
83
,
9807
9811
.
Dubois-Dalcq
,
M.
and
Armstrong
,
R.
(
1990
).
Oligodendrocyte-type 2 astrocyte lineage during myelination and remyelination
. In
Myelin: A Treatise
, (ed. R. E. Martenson)
Telford
Press
,
Caldwell
NJ.
Duncan
,
I. D.
,
Griffiths
,
I. R.
and
Munz
,
M.
(
1983
).
‘Shaking pups’: a disorder of central myelination in the spaniel dog. III. Quantitative aspects of glia and myelin in the spinal cord and optic nerve
.
Neuropath, and Applied Neurobiol
.
9
,
355
368
.
Duncan
,
I. D.
,
Hammang
,
J. P.
and
Jackson
,
K. F.
(
1987
).
Myelin mosaicism in female heterozygotes of the canine shaking pup and myelin-deficient rat mutants
.
Brain Res
.
402
,
168
172
.
Gardinier
,
M. V.
and
Macklin
,
W. B.
(
1988
).
Myelin proteolipid protein gene expression in jimpy and jimpy10 mice
.
J. Neurochem
.
51
,
360
369
.
Gencic
,
S.
,
Abuelo
,
D.
,
Ambler
,
M.
and
Hudson
,
L. D.
(
1989
).
Pelizaeus-Merzbacher disease: an X-linked neurological disorder of myelin metabolism with a novel mutation in the gene encoding proteolipid protein
.
Am. J. hum. Genet
.
45
,
435
442
.
Gencic
,
S.
and
Hudson
,
L. D.
(
1990
).
Conservative amino acid substitution in the gene encoding myelin proteolipid protein disrupts oligodendrocyte differentiation
.
J. Neurosci
.
10
,
117
124
.
Griffiths
,
I. R.
,
Duncan
,
I. D.
and
Mcculloch
,
M.
(
1981b
).
Shaking pups: a disorder of central myelination in the spaniel dog. II. Ultrastructural observations on the white matter of the cervical spinal cord
.
J. Neurocytology
10
,
847
858
.
Griffiths
,
I. R.
,
Duncan
,
I. D.
,
Mcculloch
,
M.
and
Harvey
,
M. J. A.
(
1981a
).
Shaking pups: a disorder of central myelination in the spaniel dog. I. Clinical, genetic and light-microscopical observations
.
J. Neurol. Sci
.
50
,
423
433
.
Gumpel
,
M.
,
Lachapelle
,
F.
,
Baulac
,
M.
,
Baron Van Evercooren
,
A.
,
Lubetzki
,
C.
,
Gansmuller
,
A.
,
Lombrail
,
P.
,
Jacque
,
C.
and
Baumann
,
N.
(
1987
).
Myelination in the mouse by transplanted oligodendrocytes
. In
Glial and Neuronal Communication in Development and Regeneration
, pp. 819, (
H.
Althaus
and
W.
Seifert
eds.)
Springer-Verlag, Berlin
.
Higuchi
,
R.
,
Von Beroldingen
,
C.
,
Sensabaugh
,
G.
and
Erlich
,
H.
(
1988
).
DNA typing from single hairs
.
Nature
332
,
543
546
.
Hudson
,
L. D.
,
Berndt
,
J. A.
,
Puckett
,
C.
,
Kozak
,
C. A.
and
Lazzarini
,
R. A.
(
1987
).
Aberrant splicing of proteolipid protein mRNA in the dysmyelinating jimpy mutant mouse
.
Proc. natn. Acad. Sci. U.S.A
.
84
,
1454
1458
.
Hudson
,
L. D.
,
Friedrich
,
V. L.
Jr
,
Behar
,
T.
,
Dubois-Dalcq
,
M.
and
Lazzarini
,
R. A.
(
1989b
).
The initial events in myelin synthesis: orientation of proteolipid protein in the plasma membrane of cultured oligodendrocytes
.
J. Cell Biol
.
109
,
717
727
.
Hudson
,
L. D.
and
Nadon
,
N. L.
(
1990
).
Amino acid substitutions in proteolipid protein that cause dysmyelination
. In
Myelin: A Treatise
, (ed. R. E. Martenson).
Telford
Press
,
Caldwell
NJ.
Hudson
,
L. D.
,
Puckett
,
C.
,
Berndt
,
J.
,
Chan
,
J.
and
Gencic
,
S.
(
1989a
).
Mutation of the proteolipid protein (PLP) gene in a human X-linked myelin disorder
.
Proc. natn. Acad. Sci. U S.A
.
86
,
8128
8131
.
Inuzuka
,
T.
,
Duncan
,
L. D.
and
Quarles
,
R. H.
(
1986
).
Myelin proteins in the CNS of ‘shaking pups’
.
Dev. Brain Res
.
27
,
43
50
.
Jackson
,
K. F.
and
Duncan
,
I. D.
(
1988
).
Cell kinetics and cell death in the optic nerve of the myelin deficient rat
.
J. Neurocytology
17
,
657
670
.
Kerner
,
A.-L.
and
Carson
,
J. H.
(
1984
).
Effect of the jimpy mutation on expression of myelin proteins in heterozygous and hemizygous mouse brain
.
J. Neurochem
.
43
,
1706
1715
.
Kronquist
,
K. E.
,
Crandall
,
B. E.
,
Macklin
,
W. B.
and
Campagnoni
,
A. T.
(
1987
).
Expression of myelin proteins in the developing human spinal cord: cloning and sequencing of the human proteolipid protein cDNA
.
J. Neurosa. Res
.
18
,
395
401
.
Lewis
,
S. A.
,
Balcarek
,
J. M.
,
Krek
,
V.
,
Shelanski
,
M.
and
Cowan
,
N. J.
(
1984
).
Sequence of a cDNA clone encoding mouse glial fibrillary acidic protein: structural conservation of intermediate filaments
.
Proc. natn. Acad. Sci. U.S.A
.
81
,
2743
2746
.
Macklin
,
W. B.
,
Campagnoni
,
C. W.
,
Deininger
,
P. L.
and
Gardinier
,
M. V.
(
1987b
).
Structure and expression of the mouse myelin proteolipid protein gene
.
J. Neurosci. Res
.
18
,
383
394
.
Macklin
,
W. B.
,
Gardinier
,
M. V.
,
King
,
K. D.
and
Kampf
,
K.
(
1987a
).
An AG→GG transition at a splice site in the myelin proteolipid gene in jimpy mice results in the removal of an exon
.
FEBS Lett
.
223
,
417
421
.
Maniatis
,
T.
,
Fritsch
,
E. F.
and
Sambrook
,
J.
(
1982
).
In Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor NY pp. 196
.
Matthieu
,
J.-M.
,
Widmer
,
S.
and
Herschkowttz
,
N.
(
1973
).
Jimpy, an anomaly of myelin maturation. Biochemical study of myelination phases
.
Brain Res
.
55
,
403
412
.
Meier
,
C.
and
Bischoff
,
A.
(
1974
).
Dysmyelination in jimpy mouse. Electron microscopic study
.
J. Neuropath, exp. Neurol
.
23
,
343
353
.
Nadon
,
N.
,
Arnheiter
,
H.
,
Chang
,
S.
and
Hudson
,
L.
(
1989
).
Expression of the human proteolipid protein gene in transgenic mice
.
J. cell. Biochem. 13B, 181
.
Nadon
,
N.
,
Korn
,
N.
and
Demars
,
R.
(
1988
).
A-ll: Cell type-specific and single-active-X transcription controls of newly found gene in cultured human cells
.
Som. Cell and molec. Genet
.
14
,
541
552
.
Nave
,
K.-A.
,
Bloom
,
F. E.
and
Milner
,
R. J.
(
1987b
).
A single nucleotide difference in the gene for myelin proteolipid protein defines the jimpy mutation in mouse
.
J. Neurochem
.
49
,
1873
1877
.
Nave
,
K.-A.
,
Lai
,
C.
,
Bloom
,
F. E.
and
Milner
,
R. J.
(
1987a
).
Splice site selection in the proteolipid protein (PLP) gene transcript and primary structure of the DM-20 protein of central nervous system myelin
.
Proc. natn. Acad. Sci. U.S.A
.
84
,
5665
5669
.
Puckett
,
C.
,
Hudson
,
L.
,
Ono
,
K.
,
Friedrich
,
V.
,
Benecke
,
J.
,
Dubois-Dalcq
,
M.
and
Lazzarini
,
R. A.
(
1987
).
Myelin specific proteolipid protein is expressed in myelinating Schwann cells but not incorporated into myelin sheaths
.
J. Neurosci. Res
.
18
,
511
518
.
Rackwitz
,
H.-R.
,
Zehetner
,
G.
,
Frischauf
,
A.-M.
and
Lehrach
,
H.
(
1984
).
Rapid restriction mapping of DNA cloned in lambda phage vectors
.
Gene
30
,
195
200
.
Raff
,
M. C.
(
1989
).
Glial cell diversification in the rat optic nerve
.
Science
243
,
1450
1455
.
Roussel
,
G.
,
Neskovic
,
N. M.
,
Trifilieff
,
E.
,
Artault
,
J.-C.
and
Nussbaum
,
J.-L.
(
1987
).
Arrest of proteolipid transport through the Golgi apparatus in jimpy brain
.
J. Neurocytology
10
,
195
204
.
Saiki
,
R. K.
,
Gelfand
,
D. H.
,
Stoffel
,
S.
,
Scharf
,
S.
,
Higuchi
,
R.
,
Horn
,
G. T.
,
Mulus
,
K. B.
and
Erlich
,
H. A.
(
1988
).
Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase
.
Science
239
,
487
491
.
Simons
,
R.
and
Riordan
,
J. R.
(
1990
).
The myelin deficient rat has a single base substitution in the third exon of the myelin proteolipid protein gene
.
J. Neurochem
.
54
,
1079
1081
.
Skoff
,
R. P.
(
1976
).
Myelin deficit in the jimpy mouse may be due to cellular abnormalities in astrogha
.
Nature
264
,
560
562
.
Southern
,
E. M.
(
1975
).
Detection of specific sequences among DNA fragments separated by gel electrophoresis
.
J. molec. Biol
.
98
,
503
517
.
Tabor
,
S.
and
Richardson
,
C. C.
(
1987
).
DNA sequence analysis with a modified bacteriophage T7 DNA polymerase
.
Proc. natn. Acad. Sci. U.S.A
.
84
,
4767
4771
.
Trofatter
,
J.
,
Dlouhy
,
S.
,
Demyer
,
W.
,
Conneally
,
P.
and
Hodes
,
M.
(
1989
).
Pelizaeus-Merzbacher disease: tight linkage to proteolipid protein (PLP) gene exon vanant
.
Proc. natn. Acad. Sci. U.S.A
.
86
,
9427
9430
.
Van Dorssalaer
,
A.
,
Nebhi
,
R.
,
Sorokine
,
O.
,
Schindler
,
P.
and
Luu
,
B.
(
1987
).
The DM-20 proteolipid is a major brain protein. It is synthesized earlier in fetal life than the major myelin proteolipid protein (PLP). C
.
R. Acad. Sci. Pans
305
,
555
560
.
Verity
,
A.
,
Levine
,
M.
and
Campagnoni
,
A.
(
1990
).
Cellular expression of MBP and PLP mRNAs in jimpy mice
.
Trans. Amer. Soc. Neurochem
.
1
,
236
.
Winter
,
E.
,
Yamamoto
,
F.
,
Almoguera
,
C.
and
Perucho
,
M.
(
1985
).
A method to detect and characterize point mutations in transcribed genes: amplification and overexpression of the mutant c-Ki-ras allele in human tumor cells
.
Proc. natn. Acad. Sci. U.S.A
.
82
,
7575
7579
.
Wolf
,
M. K.
,
Kardon
,
G. B.
,
Adcock
,
L. H.
and
Billings-Gagliardi
,
S.
(
1983
).
Hypomyelinated mutant mice. V. Relationship between jp and jpmld re-examined on identical genetic backgrounds
.
Brain Res
.
271
,
121
129
.
Yanagisawa
,
K.
,
Moller
,
J. R.
,
Duncan
,
I. D.
and
Quarles
,
R. H.
(
1987
).
Disproportional expression of proteolipid protein and DM-20 in the X-linked, dysmyelinating shaking pup mutant
J. Neurochem
.
49
,
1912
1917
.